Model save

README.md

@@ -18,8 +18,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3183
-- Accuracy: 0.9398
+- Loss: 0.3163
+- Accuracy: 0.9293
 
 ## Model description
 
@@ -51,10 +51,10 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss | Accuracy |
 |:-------------:|:-----:|:----:|:---------------:|:--------:|
-| 0.176         | 1.0   | 48   | 0.4412          | 0.8901   |
-| 0.0726        | 2.0   | 96   | 0.3465          | 0.9267   |
-| 0.0356        | 3.0   | 144  | 0.3606          | 0.9215   |
-| 0.0649        | 4.0   | 192  | 0.3183          | 0.9398   |
+| 0.2997        | 1.0   | 48   | 0.7613          | 0.8168   |
+| 0.1109        | 2.0   | 96   | 0.3247          | 0.9084   |
+| 0.0943        | 3.0   | 144  | 0.3684          | 0.9162   |
+| 0.0368        | 4.0   | 192  | 0.3163          | 0.9293   |
 
 
 ### Framework versions

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.0,
-    "total_flos": 3.54570337529856e+17,
-    "train_loss": 0.14949214458465576,
-    "train_runtime": 158.5609,
-    "train_samples_per_second": 28.853,
+    "epoch": 4.0,
+    "total_flos": 4.72760450039808e+17,
+    "train_loss": 0.09667936308930318,
+    "train_runtime": 211.3464,
+    "train_samples_per_second": 28.863,
     "train_steps_per_second": 0.908
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0e2bf9cde13293ce8cf990852a0f8d89daf074816d2aee7b63569bab686692ae
+oid sha256:316234cc42f7c08e247f35119eb9198b2c07c0906f0cd78fb835a067d3e63c07
 size 343267040

runs/Dec09_11-26-01_85cc5ef1c5bf/events.out.tfevents.1733744290.85cc5ef1c5bf.23.3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9956bc8358fcd30e3292d6556121e36eaa0673cc1486ad6997c041d5b60c776
+size 11335

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.0,
-    "total_flos": 3.54570337529856e+17,
-    "train_loss": 0.14949214458465576,
-    "train_runtime": 158.5609,
-    "train_samples_per_second": 28.853,
+    "epoch": 4.0,
+    "total_flos": 4.72760450039808e+17,
+    "train_loss": 0.09667936308930318,
+    "train_runtime": 211.3464,
+    "train_samples_per_second": 28.863,
     "train_steps_per_second": 0.908
 }

trainer_state.json

@@ -1,152 +1,196 @@
 {
-  "best_metric": 0.2552085518836975,
-  "best_model_checkpoint": "finetuned-bangladeshi-traditional-food/checkpoint-144",
-  "epoch": 3.0,
+  "best_metric": 0.3182551860809326,
+  "best_model_checkpoint": "finetuned-bangladeshi-traditional-food/checkpoint-192",
+  "epoch": 4.0,
   "eval_steps": 500,
-  "global_step": 144,
+  "global_step": 192,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.20833333333333334,
-      "grad_norm": 192996.96875,
-      "learning_rate": 0.00018611111111111112,
-      "loss": 0.3852,
+      "grad_norm": 119921.5,
+      "learning_rate": 0.00018958333333333332,
+      "loss": 0.2202,
       "step": 10
     },
     {
       "epoch": 0.4166666666666667,
-      "grad_norm": 121565.640625,
-      "learning_rate": 0.00017222222222222224,
-      "loss": 0.2712,
+      "grad_norm": 331074.28125,
+      "learning_rate": 0.0001791666666666667,
+      "loss": 0.1853,
       "step": 20
     },
     {
       "epoch": 0.625,
-      "grad_norm": 50398.125,
-      "learning_rate": 0.00015833333333333332,
-      "loss": 0.26,
+      "grad_norm": 264337.0,
+      "learning_rate": 0.00016875,
+      "loss": 0.2269,
       "step": 30
     },
     {
       "epoch": 0.8333333333333334,
-      "grad_norm": 54815.5703125,
-      "learning_rate": 0.00014444444444444444,
-      "loss": 0.1809,
+      "grad_norm": 103972.0625,
+      "learning_rate": 0.00015833333333333332,
+      "loss": 0.176,
       "step": 40
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.9136125654450262,
-      "eval_loss": 0.3543514907360077,
-      "eval_runtime": 6.5641,
-      "eval_samples_per_second": 58.196,
-      "eval_steps_per_second": 3.656,
+      "eval_accuracy": 0.8900523560209425,
+      "eval_loss": 0.4411921799182892,
+      "eval_runtime": 6.5407,
+      "eval_samples_per_second": 58.403,
+      "eval_steps_per_second": 3.669,
       "step": 48
     },
     {
       "epoch": 1.0416666666666667,
-      "grad_norm": 68236.78125,
-      "learning_rate": 0.00013055555555555555,
-      "loss": 0.2191,
+      "grad_norm": 7157.15185546875,
+      "learning_rate": 0.0001479166666666667,
+      "loss": 0.1162,
       "step": 50
     },
     {
       "epoch": 1.25,
-      "grad_norm": 50570.51953125,
-      "learning_rate": 0.00011666666666666668,
-      "loss": 0.1273,
+      "grad_norm": 5605.81591796875,
+      "learning_rate": 0.0001375,
+      "loss": 0.0597,
       "step": 60
     },
     {
       "epoch": 1.4583333333333333,
-      "grad_norm": 29818.142578125,
-      "learning_rate": 0.00010277777777777778,
-      "loss": 0.1002,
+      "grad_norm": 210989.8125,
+      "learning_rate": 0.00012708333333333332,
+      "loss": 0.0551,
       "step": 70
     },
     {
       "epoch": 1.6666666666666665,
-      "grad_norm": 125871.640625,
-      "learning_rate": 8.888888888888889e-05,
-      "loss": 0.177,
+      "grad_norm": 338559.40625,
+      "learning_rate": 0.00011666666666666668,
+      "loss": 0.1035,
       "step": 80
     },
     {
       "epoch": 1.875,
-      "grad_norm": 80642.8828125,
-      "learning_rate": 7.500000000000001e-05,
-      "loss": 0.0946,
+      "grad_norm": 382570.28125,
+      "learning_rate": 0.00010625000000000001,
+      "loss": 0.0726,
       "step": 90
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.9319371727748691,
-      "eval_loss": 0.29109665751457214,
-      "eval_runtime": 6.6124,
-      "eval_samples_per_second": 57.77,
-      "eval_steps_per_second": 3.63,
+      "eval_accuracy": 0.9267015706806283,
+      "eval_loss": 0.34653016924858093,
+      "eval_runtime": 6.5723,
+      "eval_samples_per_second": 58.123,
+      "eval_steps_per_second": 3.652,
       "step": 96
     },
     {
       "epoch": 2.0833333333333335,
-      "grad_norm": 7823.91357421875,
-      "learning_rate": 6.111111111111112e-05,
-      "loss": 0.0793,
+      "grad_norm": 2441.15869140625,
+      "learning_rate": 9.583333333333334e-05,
+      "loss": 0.0405,
       "step": 100
     },
     {
       "epoch": 2.2916666666666665,
-      "grad_norm": 34183.34375,
-      "learning_rate": 4.722222222222222e-05,
-      "loss": 0.0629,
+      "grad_norm": 240823.875,
+      "learning_rate": 8.541666666666666e-05,
+      "loss": 0.0559,
       "step": 110
     },
     {
       "epoch": 2.5,
-      "grad_norm": 6095.453125,
-      "learning_rate": 3.3333333333333335e-05,
-      "loss": 0.0408,
+      "grad_norm": 384660.9375,
+      "learning_rate": 7.500000000000001e-05,
+      "loss": 0.0461,
       "step": 120
     },
     {
       "epoch": 2.7083333333333335,
-      "grad_norm": 222553.59375,
-      "learning_rate": 1.9444444444444445e-05,
-      "loss": 0.076,
+      "grad_norm": 727094.5625,
+      "learning_rate": 6.458333333333334e-05,
+      "loss": 0.0519,
       "step": 130
     },
     {
       "epoch": 2.9166666666666665,
-      "grad_norm": 138554.953125,
-      "learning_rate": 5.555555555555556e-06,
-      "loss": 0.0625,
+      "grad_norm": 2871.189453125,
+      "learning_rate": 5.4166666666666664e-05,
+      "loss": 0.0356,
       "step": 140
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.9397905759162304,
-      "eval_loss": 0.2552085518836975,
-      "eval_runtime": 7.0148,
-      "eval_samples_per_second": 54.456,
-      "eval_steps_per_second": 3.421,
+      "eval_accuracy": 0.9214659685863874,
+      "eval_loss": 0.3606209456920624,
+      "eval_runtime": 6.8485,
+      "eval_samples_per_second": 55.779,
+      "eval_steps_per_second": 3.504,
       "step": 144
     },
     {
-      "epoch": 3.0,
-      "step": 144,
-      "total_flos": 3.54570337529856e+17,
-      "train_loss": 0.14949214458465576,
-      "train_runtime": 158.5609,
-      "train_samples_per_second": 28.853,
+      "epoch": 3.125,
+      "grad_norm": 1312.48974609375,
+      "learning_rate": 4.375e-05,
+      "loss": 0.0605,
+      "step": 150
+    },
+    {
+      "epoch": 3.3333333333333335,
+      "grad_norm": 251352.375,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.0467,
+      "step": 160
+    },
+    {
+      "epoch": 3.5416666666666665,
+      "grad_norm": 2361.97216796875,
+      "learning_rate": 2.2916666666666667e-05,
+      "loss": 0.1208,
+      "step": 170
+    },
+    {
+      "epoch": 3.75,
+      "grad_norm": 1161.3221435546875,
+      "learning_rate": 1.25e-05,
+      "loss": 0.078,
+      "step": 180
+    },
+    {
+      "epoch": 3.9583333333333335,
+      "grad_norm": 66789.09375,
+      "learning_rate": 2.0833333333333334e-06,
+      "loss": 0.0649,
+      "step": 190
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9397905759162304,
+      "eval_loss": 0.3182551860809326,
+      "eval_runtime": 6.6427,
+      "eval_samples_per_second": 57.507,
+      "eval_steps_per_second": 3.613,
+      "step": 192
+    },
+    {
+      "epoch": 4.0,
+      "step": 192,
+      "total_flos": 4.72760450039808e+17,
+      "train_loss": 0.09667936308930318,
+      "train_runtime": 211.3464,
+      "train_samples_per_second": 28.863,
       "train_steps_per_second": 0.908
     }
   ],
   "logging_steps": 10,
-  "max_steps": 144,
+  "max_steps": 192,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 3,
+  "num_train_epochs": 4,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -160,7 +204,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 3.54570337529856e+17,
+  "total_flos": 4.72760450039808e+17,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null